Identification of linear plants using state-space models

A method for identifying a linear plant with accessible state under servo control is described. The method is characterized by an attempt to perform the identification task on a system described by its continuous-time state-space model. Such an approach is desirable, since it allows greater insight into the physical mechanisms of the system than the input-output models. The method for identification of multiple input-output linear plants is general enough to allow for physical parametrization as well as a black-box approach to the modeling of the unknown disturbance. Initial formulation of the problem allows for identification of incremental changes rather than obsolete values of system parameters. The assumption made is that the unknown system parameters remain constant over the period of time over which identification is applied.<>